A lot of cost control effort goes into getting the number right before the job starts, then stops the moment the first fix goes in. That's backwards. The real test of a cost plan is whether anyone's still checking against it once the job's live, when actual costs start diverging from what was priced.
Cost control during construction isn't a separate discipline with its own software and its own jargon. It's the same set of numbers you priced from, looked at regularly, with someone asking whether they still hold. The contractors who do it well aren't doing anything clever. They're just doing it every month instead of once at the end.
The cost plan doesn't stop mattering once site starts
At tender, the price is a prediction. On site, it becomes a budget, and a budget is only useful if you compare it with what's really being spent.
The trouble is that site is busy. Deliveries, labour, inspections, a client who wants to change the kitchen layout. Checking numbers slides down the list because nothing seems to be on fire. Then the final account comes round and the job that looked fine turns out to have lost money in three places nobody spotted.
If the original estimate was broken down by element, measured to NRM2, you already have the structure you need. Every cost that lands on the job can be put against the element it belongs to. That's the whole trick.
Valuations as a control point, not just a payment exercise
Monthly valuations aren't just about getting paid. Done properly, they're the point where you check measured work against the cost plan and catch drift early, while there's still room to adjust.
You're already measuring what's been done to claim for it. Take that same measure and hold it against what you've spent to get there. If you've valued half the brickwork and spent two thirds of the brickwork labour budget, you know something's wrong with output, and you know it while there's still half the brickwork left to fix it on.
Treat valuations as a formality and you lose the one regular checkpoint that would have flagged a problem while it was still small.
Variations need to be priced against the plan, not in isolation
When a variation comes in, it needs pricing against the same elemental structure as the original cost plan, using NRM2 measurement, so it's clear exactly what's changed and what it's added or removed.
Say the client swaps a block and render finish for facing brick on one elevation. Priced in isolation, that's a new number on a scrap of paper. Priced against the plan, it's a clear omission of the render and block outer leaf for that area, an addition of facing brick, and a note on any change to the scaffold period. Everyone can see what moved and why.
Price variations disconnected from the original breakdown and you lose the ability to show a client or a main contractor exactly why the final account looks different from the tender figure. That's where disputes start. Your rights to payment and notice will depend on the contract you've signed, JCT, NEC or otherwise, so keep the paperwork that supports each variation from day one.
Every variation should leave a trail back to the original price. If it doesn't, it will be argued about at the final account.
A monthly routine for cost control during construction
This doesn't need a department. It needs an hour or two, the same day each month, and the discipline to do it on jobs that seem to be going well.
- Update actual costs. Invoices, labour, plant hire and subcontract payments, each put against its element.
- Compare with the valuation. For each element, what's been earned against what's been spent.
- Forecast to complete. What's left to spend on each element to finish it. This is the line that tells you where the job will land.
- Log variations and pending changes. Include instructions that haven't been priced yet, because they're still real cost.
- Write a one-page report. Elements on track, elements drifting, and what's being done about the drifting ones.
Cost reports only work if they're built to be read by whoever's running the job, not just filed. A report that flags where the job's tracking against plan, and where it isn't, is worth far more than a long document nobody opens until the final account.
Where control usually slips
The two things that undo cost control on site are the same two that undo cost planning before it: scope that shifts without anyone updating the numbers, and a gap between the office and the site that means nobody's comparing plan to actual until it's too late.
There are smaller leaks too. Day-work sheets that don't get signed. Materials ordered for one job and used on another. Preliminaries that run on for weeks after the programme said they'd finish, with nobody noticing that the welfare cabin is still on hire. Each one is small. Together they're a margin. For more ways to stop them, see keeping a job's cost under control once you're on site.
Neither the big failures nor the small ones are complicated to fix. Both require someone actually doing the comparison every month, not just at the end.
One number to watch
If you only look at one figure each month, make it the forecast final cost for each element against its budget. Spend to date can look fine on a job that's heading for a loss. Forecast to complete can't hide it for long. That's cost control during construction in one line: know where each element will finish, not just where it is today.
If you want a properly structured base to control against, our cost planning services build the plan from your drawings, measured to NRM2 and priced on current rates local to the job, so every element has a number worth tracking.